The goal of this project is to establish a novel paradigm of dense angularly sampled and sparse intensity- modulated radiation therapy (DASSIM-RT). In this scheme, the redundant or dispensable modulation of the incident intensity-modulated beams is removed effectively by using a compressed sensing (CS) technique. The delivery time saved in this way is used to increase the angular sampling for improved dose conformality. By balancing the angular sampling and intensity modulation, DASSIM-RT enables us to fully utilize the technical capabilities of modern digital linacs to produce highly conformal dose distributions that can be delivered efficiently. Specifically, we wil 1) set up a compressed sensing (CS)-based framework for inverse treatment planning; 2) investigate a new type of treatment scheme termed DASSIM-RT; and 3) show the advantage of DASSIM-RT through a series of phantom cases and previously treated patients. DASSIM-RT represents a truly optimal RT scheme with uncompromised angular sampling (including non-coplanar beams), beam intensity modulation, and possible field-specific energy and collimator angle. If successful, the project will allow us to overcome many of the limitations of existing treatment schemes to meet the unmet clinical demand for highly conformal dose distributions in radiation oncology.